In Maryland, an artificial intelligence (AI) security system mistakenly identified a bag of chips as a firearm, resulting in the temporary arrest of a high school student.
The incident occurred at Kenwood Middle School when student Taki Allen was walking through the corridor holding a bag of Doritos. The AI threat detection system flagged the snack as a potential weapon. School security personnel quickly responded, restraining the student and placing him in handcuffs before realizing the error.
School principal Katie Smith confirmed that the alert was promptly canceled by the security team, but she was not notified in time. As a result, a school resource officer was called, who then involved local police.
Omnilert, the company behind the AI system, expressed regret over the incident and offered support to the student and his family. The company emphasized that the system operated according to its programmed procedures and is designed to minimize human error in threat detection.
Representatives from Omnilert noted that the technology continuously learns from new data, and incidents like this contribute to improving the algorithms. Their AI security systems are deployed in over a hundred schools across the U.S., where they have successfully prevented real threats in the past.
Allen’s parents reported that the event caused significant stress for their son and raised concerns about the safety of using such AI systems in schools. The local education authorities have launched an internal review to assess whether staff followed the proper protocols.
This incident follows other recent cases highlighting AI-related risks in schools, including the detention of a Florida student over questions posed to ChatGPT about harming a friend.
